214 DISCOVERY REPORTS 



seem to be due to the rapid recession, in comparatively recent times, of the glacier 

 which formerly filled it. A somewhat similar condition may be seen in the hinterland 

 of this portion of the island, where some of the glaciers have receded leaving moraine- 

 dammed lakes. 



As will be seen from Table LV, the salinity and phosphate values in the upper layers 

 of Moranen Fjord were low, while deep down they were high and the temperatures 

 very low, minus temperatures being found at depths of 75 and 100 m. as late as 



January 22. 



Water obtained by melting a piece of very black morainic ice appeared to contain 

 only 40 mg. P2O5 per m.^ as against the 54 mg. of the surface water of Moranen Fjord, 

 so that the diminution of phosphate in the surface waters of Cumberland Bay by land 

 drainage becomes readily understandable. 



The water at St. MS 94 was rendered turbid by melting morainic ice, yet a heavy 

 catch of Coscinodiscus stellaris was obtained, the conditions being entirely different from 

 those at the routine station in the bay a week later, when the same species again occurred 

 in bulk. A possible explanation is that some internal biological factor determines the 

 period for its maximum, within narrow limits during the month of January, and that 

 being everywhere present inshore in small numbers, it so happened that the maximum 

 in Moranen Fjord became apparent earlier than in the bay where conditions at the time 

 were less favourable. That Coscinodisais stellaris is nearly always present in inshore waters 

 around South Georgia could be seen from the samples taken at the innermost stations 

 on the Unes worked by the R.R.S. 'Discovery 11', but conditions in Cumberland Bay 

 on January 8, the day of the Moranen Fjord maximum, were unstable, and the outburst 

 of Coscinodiscus due to outside invasion came a week later. A more probable explanation, 

 therefore, is that being a form with a heavily silicified skeleton it requires fairly stable 

 conditions as regards density in order to develop in large numbers. At the station in 

 Moranen Fjord it was present in large numbers belozv 35 m., and since there appeared 

 to be a sharp discontinuity layer between 40 and 50 m. it may be that this prevented 

 the heavy frustules from sinking farther. Many of those obtained in the deeper haul, 

 50-0 m., were dead. Its abundance in the bay with the invasion of outside water of 

 much greater density a week later, was confined almost entirely to the upper 25 m., and 

 here (see Station List, St. MS 95) the most marked difference in density occurred 



between 20 and 30 m. 



It thus appears probable that Coscinodiscus stellaris is a form requiring stable con- 

 ditions as regards density in order to attain its maximum development, that a sharp 

 discontinuity layer tending to keep its heav}' frustules within the zone of optimum light 

 intensity is necessary, and that in the absence of such layering the frustules sink rapidly. 

 It is also probable that mid-season is the normal time for this species to attain its 

 maximum, as with the thawing of the lower slopes just complete, it is at that time that 

 the most marked layering is liable to occur, always provided that wind conditions render 

 this possible. In this connection it is interesting to note that there is evidence that 

 other southern members of the genus Coscitiodiscus, notably C. bouvet, show a strong 



